Introduction¶

  • Aim of this notebook is to run Vanilla CNN and Fine-Tune VGG16 - for Dogs and Cats Classification using tensorflow.

Obtaining the Data¶

Load the images folders & required libraries¶

In [ ]:
from tensorflow import keras
from tensorflow.keras import layers
from keras.callbacks import ModelCheckpoint
import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
import seaborn as sns
import pandas as pd
import warnings
warnings.filterwarnings('ignore')

import os, shutil, pathlib

train_dir = pathlib.Path("data/train")
new_base_dir = pathlib.Path("datacats_and_dogs_small")


def make_subset(subset_name, start_index, end_index):
    for category in ("cat", "dog"):
        dir = new_base_dir / subset_name / category
        os.makedirs(dir)
        fnames = [f"{category}.{i}.jpg" for i in range(start_index, end_index)]
        for fname in fnames:
            shutil.copyfile(src=train_dir / fname,
                            dst=dir / fname)

make_subset("train", start_index=0, end_index=1000)
make_subset("validation", start_index=1000, end_index=1500)
make_subset("test", start_index=1500, end_index=2500)

Create Datasets for Images¶

In [ ]:
from tensorflow.keras.utils import image_dataset_from_directory


data_folder = pathlib.Path('datacats_and_dogs_small')


train_dataset = image_dataset_from_directory(
    data_folder / "train",
    image_size=(180, 180),
    batch_size=32)
validation_dataset = image_dataset_from_directory(
    data_folder / "validation",
    image_size=(180, 180),
    batch_size=32)
test_dataset = image_dataset_from_directory(
    data_folder / "test",
    image_size=(180, 180),
    batch_size=32)

Found 2000 files belonging to 2 classes.
Found 1000 files belonging to 2 classes.
Found 2000 files belonging to 2 classes.
In [ ]:
type(train_dataset)
Out[ ]:
tensorflow.python.data.ops.batch_op._BatchDataset
In [ ]:
for data_batch, labels_batch in train_dataset:
    print("data batch shape:", data_batch.shape)
    print("labels batch shape:", labels_batch.shape)
    break

data batch shape: (32, 180, 180, 3)
labels batch shape: (32,)
In [ ]:
import matplotlib.pyplot as plt

plt.imshow(data_batch[0].numpy().astype("uint8"))
Out[ ]:
<matplotlib.image.AxesImage at 0x1e17324e750>

Exploratory Data Analysis¶

  • Aim of this section is to explore the dataset in some depth. This includes looking at some examples of each image, exploring the distribution of cat vs dog classes, and studying the image dimensions.

Display First 10 Cats & Dogs Images¶

In [ ]:
import matplotlib.pyplot as plt
import matplotlib.image as mpimg

cols = 5
rows = 2  
fig = plt.figure(figsize=(4 * cols - 1, 4.5 * rows - 1))
for i in range(1, 11):
    img_path = f'datacats_and_dogs_small/train/cat/cat.{i}.jpg'  
    img = mpimg.imread(img_path)
    ax = fig.add_subplot(rows, cols, i)
    ax.imshow(img)
    ax.set_title(f'Cat {i}')
    ax.axis('off') 
fig.suptitle('First 10 Cat Images', fontsize=16)  
plt.tight_layout() 
plt.show()
plt.close()

fig = plt.figure(figsize=(4 * cols - 1, 4.5 * rows - 1))
for i in range(1, 11):
    img_path = f'datacats_and_dogs_small/train/dog/dog.{i}.jpg'  
    img = mpimg.imread(img_path)
    ax = fig.add_subplot(rows, cols, i)
    ax.imshow(img)
    ax.set_title(f'Cat {i}')
    ax.axis('off') 
fig.suptitle('First 10 Dog Images', fontsize=16)  
plt.tight_layout() 
plt.show()
plt.close()

Frequency of Cats & Dogs in Training Set¶

In [ ]:
total_cats = len(os.listdir('datacats_and_dogs_small/train/cat'))
total_dogs = len(os.listdir('datacats_and_dogs_small/train/dog'))

pd.DataFrame({'category': ['cat', 'dog'], 'count': [total_cats, total_dogs]}).plot.bar(x='category', y='count', legend=False)
Out[ ]:
<Axes: xlabel='category'>
  • Both Cats & Dogs have equal number images which indicates a balanced dataset.

Image Dimensions¶

In [ ]:
# function to get image dimensions
def get_image_dimensions(folder_path):
    image_dimensions = []
    for filename in os.listdir(folder_path):
        if filename.endswith('.jpg') or filename.endswith('.jpg'):
            image_path = os.path.join(folder_path, filename)
            img = mpimg.imread(image_path)
            if img is not None:  
                height, width, _ = img.shape
                image_dimensions.append((filename, width, height))
    return image_dimensions

cat_folder_path = 'datacats_and_dogs_small/train/cat'
dog_folder_path = 'datacats_and_dogs_small/train/dog'
# tuples with dimensions of images
cat_dimensions = get_image_dimensions(cat_folder_path)
dog_dimensions = get_image_dimensions(dog_folder_path)

# Convert the list of tuples to a DataFrame
cat_df = pd.DataFrame(cat_dimensions, columns=['filename', 'width', 'height'])
dog_df = pd.DataFrame(dog_dimensions, columns=['filename', 'width', 'height'])

1. Distribution of Heights¶

In [ ]:
combined_height_df = pd.concat([
    cat_df[['height']].assign(image_type='cat'),
    dog_df[['height']].assign(image_type='dog')
])

fig = plt.figure(figsize=(12, 6))
sns.displot(data=combined_height_df, x='height', hue='image_type', kde=True,col='image_type', palette='viridis',legend=False)
plt.suptitle('Distribution of Heights of Cat and Dog Images', x=0.5, y=1.05, ha='center', fontsize=16) 
plt.show()

<Figure size 1200x600 with 0 Axes>

2. Distribution of Width¶

In [ ]:
combined_width_df = pd.concat([
    cat_df[['width']].assign(image_type='cat'),
    dog_df[['width']].assign(image_type='dog')
])

fig = plt.figure(figsize=(12, 6))
sns.displot(data=combined_width_df, x='width', hue='image_type', kde=True,col='image_type',legend=False)
plt.suptitle('Distribution of Widths of Cat and Dog Images', x=0.5, y=1.05, ha='center', fontsize=16) 
plt.show()

<Figure size 1200x600 with 0 Axes>

Model Creation¶

Model 1¶

In [ ]:
inputs = keras.Input(shape=(180, 180, 3))
x = layers.Rescaling(1./255)(inputs)
x = layers.Conv2D(filters=32, kernel_size=3, activation="relu")(x)
x = layers.MaxPooling2D(pool_size=2)(x)
x = layers.Conv2D(filters=64, kernel_size=3, activation="relu")(x)
x = layers.MaxPooling2D(pool_size=2)(x)
x = layers.Conv2D(filters=128, kernel_size=3, activation="relu")(x)
x = layers.MaxPooling2D(pool_size=2)(x)
x = layers.Conv2D(filters=256, kernel_size=3, activation="relu")(x)
x = layers.MaxPooling2D(pool_size=2)(x)
x = layers.Conv2D(filters=256, kernel_size=3, activation="relu")(x)
x = layers.Flatten()(x)
outputs = layers.Dense(1, activation="sigmoid")(x)
model_1 = keras.Model(inputs=inputs, outputs=outputs)
In [ ]:
model_1.summary()

Model: "model"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_1 (InputLayer)        [(None, 180, 180, 3)]     0         
                                                                 
 rescaling (Rescaling)       (None, 180, 180, 3)       0         
                                                                 
 conv2d (Conv2D)             (None, 178, 178, 32)      896       
                                                                 
 max_pooling2d (MaxPooling2D  (None, 89, 89, 32)       0         
 )                                                               
                                                                 
 conv2d_1 (Conv2D)           (None, 87, 87, 64)        18496     
                                                                 
 max_pooling2d_1 (MaxPooling  (None, 43, 43, 64)       0         
 2D)                                                             
                                                                 
 conv2d_2 (Conv2D)           (None, 41, 41, 128)       73856     
                                                                 
 max_pooling2d_2 (MaxPooling  (None, 20, 20, 128)      0         
 2D)                                                             
                                                                 
 conv2d_3 (Conv2D)           (None, 18, 18, 256)       295168    
                                                                 
 max_pooling2d_3 (MaxPooling  (None, 9, 9, 256)        0         
 2D)                                                             
                                                                 
 conv2d_4 (Conv2D)           (None, 7, 7, 256)         590080    
                                                                 
 flatten (Flatten)           (None, 12544)             0         
                                                                 
 dense (Dense)               (None, 1)                 12545     
                                                                 
=================================================================
Total params: 991,041
Trainable params: 991,041
Non-trainable params: 0
_________________________________________________________________
In [ ]:
model_1.compile(loss="binary_crossentropy",
              optimizer="rmsprop",
              metrics=["accuracy"])
In [ ]:
checkpoint_filepath = 'best_model_model1_copy.hdf5' 

callbacks = [
    keras.callbacks.ModelCheckpoint(
        filepath=checkpoint_filepath,
        save_best_only=True,
        monitor="val_loss",
        mode='min')
]
history_1 = model_1.fit(
    train_dataset,
    epochs=30,
    validation_data=validation_dataset,
    callbacks=callbacks)

Epoch 1/30
63/63 [==============================] - 46s 727ms/step - loss: 0.7130 - accuracy: 0.5160 - val_loss: 0.6921 - val_accuracy: 0.5860
Epoch 2/30
63/63 [==============================] - 52s 820ms/step - loss: 0.6966 - accuracy: 0.5235 - val_loss: 0.7221 - val_accuracy: 0.5000
Epoch 3/30
63/63 [==============================] - 47s 748ms/step - loss: 0.6847 - accuracy: 0.5655 - val_loss: 0.6552 - val_accuracy: 0.6110
Epoch 4/30
63/63 [==============================] - 48s 761ms/step - loss: 0.6405 - accuracy: 0.6550 - val_loss: 0.6578 - val_accuracy: 0.6410
Epoch 5/30
63/63 [==============================] - 49s 770ms/step - loss: 0.6075 - accuracy: 0.6565 - val_loss: 0.6766 - val_accuracy: 0.6060
Epoch 6/30
63/63 [==============================] - 50s 796ms/step - loss: 0.5756 - accuracy: 0.6925 - val_loss: 0.6132 - val_accuracy: 0.6610
Epoch 7/30
63/63 [==============================] - 53s 833ms/step - loss: 0.5604 - accuracy: 0.7160 - val_loss: 0.6197 - val_accuracy: 0.6760
Epoch 8/30
63/63 [==============================] - 54s 863ms/step - loss: 0.5158 - accuracy: 0.7350 - val_loss: 0.5874 - val_accuracy: 0.6930
Epoch 9/30
63/63 [==============================] - 53s 834ms/step - loss: 0.4842 - accuracy: 0.7690 - val_loss: 0.6208 - val_accuracy: 0.6990
Epoch 10/30
63/63 [==============================] - 52s 830ms/step - loss: 0.4552 - accuracy: 0.7905 - val_loss: 0.7135 - val_accuracy: 0.6600
Epoch 11/30
63/63 [==============================] - 52s 827ms/step - loss: 0.4195 - accuracy: 0.8135 - val_loss: 0.7523 - val_accuracy: 0.7010
Epoch 12/30
63/63 [==============================] - 53s 834ms/step - loss: 0.3730 - accuracy: 0.8375 - val_loss: 0.7074 - val_accuracy: 0.7120
Epoch 13/30
63/63 [==============================] - 52s 831ms/step - loss: 0.3182 - accuracy: 0.8645 - val_loss: 0.9930 - val_accuracy: 0.6320
Epoch 14/30
63/63 [==============================] - 53s 837ms/step - loss: 0.2967 - accuracy: 0.8775 - val_loss: 0.7207 - val_accuracy: 0.7410
Epoch 15/30
63/63 [==============================] - 53s 838ms/step - loss: 0.2448 - accuracy: 0.9000 - val_loss: 0.8259 - val_accuracy: 0.7230
Epoch 16/30
63/63 [==============================] - 53s 840ms/step - loss: 0.1753 - accuracy: 0.9285 - val_loss: 1.1771 - val_accuracy: 0.7070
Epoch 17/30
63/63 [==============================] - 53s 840ms/step - loss: 0.1617 - accuracy: 0.9380 - val_loss: 0.9725 - val_accuracy: 0.7290
Epoch 18/30
63/63 [==============================] - 53s 847ms/step - loss: 0.1302 - accuracy: 0.9550 - val_loss: 1.1970 - val_accuracy: 0.7380
Epoch 19/30
63/63 [==============================] - 53s 837ms/step - loss: 0.1040 - accuracy: 0.9620 - val_loss: 1.4788 - val_accuracy: 0.7050
Epoch 20/30
63/63 [==============================] - 53s 841ms/step - loss: 0.1007 - accuracy: 0.9635 - val_loss: 1.1852 - val_accuracy: 0.7470
Epoch 21/30
63/63 [==============================] - 53s 842ms/step - loss: 0.0749 - accuracy: 0.9740 - val_loss: 1.6978 - val_accuracy: 0.6810
Epoch 22/30
63/63 [==============================] - 53s 836ms/step - loss: 0.0586 - accuracy: 0.9785 - val_loss: 1.5462 - val_accuracy: 0.7260
Epoch 23/30
63/63 [==============================] - 53s 842ms/step - loss: 0.0701 - accuracy: 0.9800 - val_loss: 1.6727 - val_accuracy: 0.7370
Epoch 24/30
63/63 [==============================] - 53s 839ms/step - loss: 0.0802 - accuracy: 0.9715 - val_loss: 1.4721 - val_accuracy: 0.7450
Epoch 25/30
63/63 [==============================] - 53s 847ms/step - loss: 0.0615 - accuracy: 0.9775 - val_loss: 1.4910 - val_accuracy: 0.7370
Epoch 26/30
63/63 [==============================] - 53s 844ms/step - loss: 0.0369 - accuracy: 0.9885 - val_loss: 2.0298 - val_accuracy: 0.7050
Epoch 27/30
63/63 [==============================] - 53s 839ms/step - loss: 0.0483 - accuracy: 0.9835 - val_loss: 1.7366 - val_accuracy: 0.7390
Epoch 28/30
63/63 [==============================] - 53s 848ms/step - loss: 0.0500 - accuracy: 0.9815 - val_loss: 2.3627 - val_accuracy: 0.7080
Epoch 29/30
63/63 [==============================] - 53s 837ms/step - loss: 0.0546 - accuracy: 0.9855 - val_loss: 2.7954 - val_accuracy: 0.6700
Epoch 30/30
63/63 [==============================] - 53s 833ms/step - loss: 0.0690 - accuracy: 0.9785 - val_loss: 2.0654 - val_accuracy: 0.7150

Plots for Evaluating the model¶

In [ ]:
accuracy = history_1.history["accuracy"]
val_accuracy = history_1.history["val_accuracy"]
loss = history_1.history["loss"]
val_loss = history_1.history["val_loss"]
epochs = range(1, len(accuracy) + 1)
plt.plot(epochs, accuracy, "bo", label="Training accuracy")
plt.plot(epochs, val_accuracy, "b", label="Validation accuracy")
plt.title("Training and validation accuracy")
plt.legend()
plt.figure()
plt.plot(epochs, loss, "bo", label="Training loss")
plt.plot(epochs, val_loss, "b", label="Validation loss")
plt.title("Training and validation loss")
plt.legend()
plt.show()

Testing the model with test set¶

In [ ]:
test_model_1 = keras.models.load_model(
    "best_model_model1.hdf5")
test_loss_1, test_acc_1 = test_model_1.evaluate(test_dataset)
print(f"Test accuracy: {test_acc_1:.3f}")

63/63 [==============================] - 14s 219ms/step - loss: 1.1131 - accuracy: 0.7290
Test accuracy: 0.729
In [ ]:
history_1_dict = history_1.history
hist_df_1 = pd.DataFrame(history_1_dict)
hist_df_1
Out[ ]:
loss accuracy val_loss val_accuracy
0 0.713015 0.5160 0.692076 0.586
1 0.696564 0.5235 0.722122 0.500
2 0.684678 0.5655 0.655198 0.611
3 0.640484 0.6550 0.657761 0.641
4 0.607469 0.6565 0.676594 0.606
5 0.575559 0.6925 0.613168 0.661
6 0.560372 0.7160 0.619738 0.676
7 0.515783 0.7350 0.587357 0.693
8 0.484195 0.7690 0.620832 0.699
9 0.455219 0.7905 0.713530 0.660
10 0.419528 0.8135 0.752285 0.701
11 0.372983 0.8375 0.707385 0.712
12 0.318205 0.8645 0.992980 0.632
13 0.296675 0.8775 0.720664 0.741
14 0.244807 0.9000 0.825889 0.723
15 0.175319 0.9285 1.177084 0.707
16 0.161705 0.9380 0.972517 0.729
17 0.130194 0.9550 1.196969 0.738
18 0.104024 0.9620 1.478766 0.705
19 0.100650 0.9635 1.185195 0.747
20 0.074881 0.9740 1.697806 0.681
21 0.058640 0.9785 1.546249 0.726
22 0.070146 0.9800 1.672654 0.737
23 0.080200 0.9715 1.472115 0.745
24 0.061488 0.9775 1.490953 0.737
25 0.036924 0.9885 2.029843 0.705
26 0.048272 0.9835 1.736598 0.739
27 0.050037 0.9815 2.362741 0.708
28 0.054592 0.9855 2.795363 0.670
29 0.068986 0.9785 2.065430 0.715

Conclusions¶

Training & Validation loss

  • This model performs well on the training data but not on validation data.
  • The validation loss start to increase after epoch 10. The model have less loss until the epoch 10.
  • This means that the model was trying to stabilize to a point till epoch 10.
  • The gap between validation and training loss is very high.
  • This is an indicator of overfitting.

Accuracy

  • We are getting good accuracy for training set.
  • But for validation accuracy, it has started to keep on same value range after epoch 10.
  • For test set, we are getting accuracy of .729. This means that 72.9 % of images are predicted correctly by the model.

Weights

  • The best set of weights seems to correspond to epoch 13, where val_accuracy is the highest (0.741) and val_loss is relatively low compared to other epochs (0.720664).

Model 2 - Fine Tuning with vgg16¶

In [ ]:
conv_base = keras.applications.vgg16.VGG16(
    weights="imagenet",
    include_top=False,
    input_shape=(180, 180, 3))
In [ ]:
conv_base.summary()

Model: "vgg16"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_2 (InputLayer)        [(None, 180, 180, 3)]     0         
                                                                 
 block1_conv1 (Conv2D)       (None, 180, 180, 64)      1792      
                                                                 
 block1_conv2 (Conv2D)       (None, 180, 180, 64)      36928     
                                                                 
 block1_pool (MaxPooling2D)  (None, 90, 90, 64)        0         
                                                                 
 block2_conv1 (Conv2D)       (None, 90, 90, 128)       73856     
                                                                 
 block2_conv2 (Conv2D)       (None, 90, 90, 128)       147584    
                                                                 
 block2_pool (MaxPooling2D)  (None, 45, 45, 128)       0         
                                                                 
 block3_conv1 (Conv2D)       (None, 45, 45, 256)       295168    
                                                                 
 block3_conv2 (Conv2D)       (None, 45, 45, 256)       590080    
                                                                 
 block3_conv3 (Conv2D)       (None, 45, 45, 256)       590080    
                                                                 
 block3_pool (MaxPooling2D)  (None, 22, 22, 256)       0         
                                                                 
 block4_conv1 (Conv2D)       (None, 22, 22, 512)       1180160   
                                                                 
 block4_conv2 (Conv2D)       (None, 22, 22, 512)       2359808   
                                                                 
 block4_conv3 (Conv2D)       (None, 22, 22, 512)       2359808   
                                                                 
 block4_pool (MaxPooling2D)  (None, 11, 11, 512)       0         
                                                                 
 block5_conv1 (Conv2D)       (None, 11, 11, 512)       2359808   
                                                                 
 block5_conv2 (Conv2D)       (None, 11, 11, 512)       2359808   
                                                                 
 block5_conv3 (Conv2D)       (None, 11, 11, 512)       2359808   
                                                                 
 block5_pool (MaxPooling2D)  (None, 5, 5, 512)         0         
                                                                 
=================================================================
Total params: 14,714,688
Trainable params: 14,714,688
Non-trainable params: 0
_________________________________________________________________

Preprocessing the input¶

In [ ]:
import numpy as np

def get_features_and_labels(dataset):
    all_features = []
    all_labels = []
    for images, labels in dataset:
        preprocessed_images = keras.applications.vgg16.preprocess_input(images)
        features = conv_base.predict(preprocessed_images)
        all_features.append(features)
        all_labels.append(labels)
    return np.concatenate(all_features), np.concatenate(all_labels)

train_features, train_labels =  get_features_and_labels(train_dataset)
val_features, val_labels =  get_features_and_labels(validation_dataset)
test_features, test_labels =  get_features_and_labels(test_dataset)

1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 2s 2s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 1s 704ms/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 3s 3s/step
1/1 [==============================] - 1s 1s/step

Freezing the model weights¶

In [ ]:
conv_base.trainable = False
conv_base.summary(
)

Model: "vgg16"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_2 (InputLayer)        [(None, 180, 180, 3)]     0         
                                                                 
 block1_conv1 (Conv2D)       (None, 180, 180, 64)      1792      
                                                                 
 block1_conv2 (Conv2D)       (None, 180, 180, 64)      36928     
                                                                 
 block1_pool (MaxPooling2D)  (None, 90, 90, 64)        0         
                                                                 
 block2_conv1 (Conv2D)       (None, 90, 90, 128)       73856     
                                                                 
 block2_conv2 (Conv2D)       (None, 90, 90, 128)       147584    
                                                                 
 block2_pool (MaxPooling2D)  (None, 45, 45, 128)       0         
                                                                 
 block3_conv1 (Conv2D)       (None, 45, 45, 256)       295168    
                                                                 
 block3_conv2 (Conv2D)       (None, 45, 45, 256)       590080    
                                                                 
 block3_conv3 (Conv2D)       (None, 45, 45, 256)       590080    
                                                                 
 block3_pool (MaxPooling2D)  (None, 22, 22, 256)       0         
                                                                 
 block4_conv1 (Conv2D)       (None, 22, 22, 512)       1180160   
                                                                 
 block4_conv2 (Conv2D)       (None, 22, 22, 512)       2359808   
                                                                 
 block4_conv3 (Conv2D)       (None, 22, 22, 512)       2359808   
                                                                 
 block4_pool (MaxPooling2D)  (None, 11, 11, 512)       0         
                                                                 
 block5_conv1 (Conv2D)       (None, 11, 11, 512)       2359808   
                                                                 
 block5_conv2 (Conv2D)       (None, 11, 11, 512)       2359808   
                                                                 
 block5_conv3 (Conv2D)       (None, 11, 11, 512)       2359808   
                                                                 
 block5_pool (MaxPooling2D)  (None, 5, 5, 512)         0         
                                                                 
=================================================================
Total params: 14,714,688
Trainable params: 0
Non-trainable params: 14,714,688
_________________________________________________________________

Create Model with Convolutional base¶

In [ ]:
data_augmentation = keras.Sequential(
    [
        layers.RandomFlip("horizontal"),
        layers.RandomRotation(0.1),
        layers.RandomZoom(0.2),
    ]
)

inputs = keras.Input(shape=(180, 180, 3))
x = data_augmentation(inputs)
x = keras.applications.vgg16.preprocess_input(x)
x = conv_base(x)
x = layers.Flatten()(x)
x = layers.Dense(256)(x)
x = layers.Dropout(0.5)(x)
outputs = layers.Dense(1, activation="sigmoid")(x)
model_2 = keras.Model(inputs, outputs)
In [ ]:
model_2.summary()

Model: "model_1"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_3 (InputLayer)        [(None, 180, 180, 3)]     0         
                                                                 
 sequential (Sequential)     (None, 180, 180, 3)       0         
                                                                 
 tf.__operators__.getitem (S  (None, 180, 180, 3)      0         
 licingOpLambda)                                                 
                                                                 
 tf.nn.bias_add (TFOpLambda)  (None, 180, 180, 3)      0         
                                                                 
 vgg16 (Functional)          (None, 5, 5, 512)         14714688  
                                                                 
 flatten_1 (Flatten)         (None, 12800)             0         
                                                                 
 dense_1 (Dense)             (None, 256)               3277056   
                                                                 
 dropout (Dropout)           (None, 256)               0         
                                                                 
 dense_2 (Dense)             (None, 1)                 257       
                                                                 
=================================================================
Total params: 17,992,001
Trainable params: 3,277,313
Non-trainable params: 14,714,688
_________________________________________________________________
In [ ]:
model_2.compile(loss="binary_crossentropy",
              optimizer="rmsprop",
              metrics=["accuracy"])


checkpoint_filepath = 'best_model_copy.hdf5' 

callbacks = [
    keras.callbacks.ModelCheckpoint(
        filepath=checkpoint_filepath,
        save_best_only=True,
        monitor="val_loss",
        mode='min')]
In [ ]:
history_2 = model_2.fit(
    train_dataset,
    epochs=50,
    validation_data=validation_dataset,
    callbacks=callbacks)

Epoch 1/50
63/63 [==============================] - 280s 4s/step - loss: 17.1685 - accuracy: 0.8895 - val_loss: 4.9841 - val_accuracy: 0.9650
Epoch 2/50
63/63 [==============================] - 275s 4s/step - loss: 6.1005 - accuracy: 0.9440 - val_loss: 3.5135 - val_accuracy: 0.9740
Epoch 3/50
63/63 [==============================] - 274s 4s/step - loss: 4.5690 - accuracy: 0.9555 - val_loss: 7.9375 - val_accuracy: 0.9560
Epoch 4/50
63/63 [==============================] - 275s 4s/step - loss: 4.8805 - accuracy: 0.9600 - val_loss: 3.7878 - val_accuracy: 0.9780
Epoch 5/50
63/63 [==============================] - 275s 4s/step - loss: 4.3384 - accuracy: 0.9570 - val_loss: 3.4223 - val_accuracy: 0.9740
Epoch 6/50
63/63 [==============================] - 278s 4s/step - loss: 3.0382 - accuracy: 0.9710 - val_loss: 5.0258 - val_accuracy: 0.9700
Epoch 7/50
63/63 [==============================] - 277s 4s/step - loss: 4.6583 - accuracy: 0.9675 - val_loss: 7.0391 - val_accuracy: 0.9630
Epoch 8/50
63/63 [==============================] - 276s 4s/step - loss: 2.9847 - accuracy: 0.9710 - val_loss: 3.8607 - val_accuracy: 0.9720
Epoch 9/50
63/63 [==============================] - 274s 4s/step - loss: 3.2730 - accuracy: 0.9730 - val_loss: 4.5456 - val_accuracy: 0.9700
Epoch 10/50
63/63 [==============================] - 277s 4s/step - loss: 1.8048 - accuracy: 0.9785 - val_loss: 4.7965 - val_accuracy: 0.9720
Epoch 11/50
63/63 [==============================] - 276s 4s/step - loss: 2.5690 - accuracy: 0.9740 - val_loss: 2.5217 - val_accuracy: 0.9760
Epoch 12/50
63/63 [==============================] - 276s 4s/step - loss: 1.7968 - accuracy: 0.9795 - val_loss: 3.0949 - val_accuracy: 0.9750
Epoch 13/50
63/63 [==============================] - 280s 4s/step - loss: 2.5654 - accuracy: 0.9725 - val_loss: 5.0855 - val_accuracy: 0.9630
Epoch 14/50
63/63 [==============================] - 277s 4s/step - loss: 1.6234 - accuracy: 0.9785 - val_loss: 4.5574 - val_accuracy: 0.9660
Epoch 15/50
63/63 [==============================] - 276s 4s/step - loss: 1.9424 - accuracy: 0.9750 - val_loss: 2.9817 - val_accuracy: 0.9790
Epoch 16/50
63/63 [==============================] - 277s 4s/step - loss: 1.5701 - accuracy: 0.9820 - val_loss: 2.9405 - val_accuracy: 0.9770
Epoch 17/50
63/63 [==============================] - 277s 4s/step - loss: 0.9354 - accuracy: 0.9845 - val_loss: 4.1686 - val_accuracy: 0.9760
Epoch 18/50
63/63 [==============================] - 278s 4s/step - loss: 1.0732 - accuracy: 0.9855 - val_loss: 3.5293 - val_accuracy: 0.9770
Epoch 19/50
63/63 [==============================] - 278s 4s/step - loss: 1.6693 - accuracy: 0.9810 - val_loss: 3.3674 - val_accuracy: 0.9740
Epoch 20/50
63/63 [==============================] - 277s 4s/step - loss: 1.3279 - accuracy: 0.9820 - val_loss: 2.7475 - val_accuracy: 0.9750
Epoch 21/50
63/63 [==============================] - 276s 4s/step - loss: 1.3281 - accuracy: 0.9800 - val_loss: 2.4419 - val_accuracy: 0.9750
Epoch 22/50
63/63 [==============================] - 277s 4s/step - loss: 1.0107 - accuracy: 0.9845 - val_loss: 2.4627 - val_accuracy: 0.9760
Epoch 23/50
63/63 [==============================] - 276s 4s/step - loss: 0.6847 - accuracy: 0.9865 - val_loss: 5.7680 - val_accuracy: 0.9550
Epoch 24/50
63/63 [==============================] - 276s 4s/step - loss: 1.0973 - accuracy: 0.9825 - val_loss: 2.8988 - val_accuracy: 0.9760
Epoch 25/50
63/63 [==============================] - 278s 4s/step - loss: 1.1435 - accuracy: 0.9825 - val_loss: 2.7116 - val_accuracy: 0.9750
Epoch 26/50
63/63 [==============================] - 278s 4s/step - loss: 0.8626 - accuracy: 0.9800 - val_loss: 2.4165 - val_accuracy: 0.9770
Epoch 27/50
63/63 [==============================] - 281s 4s/step - loss: 0.8594 - accuracy: 0.9845 - val_loss: 2.2589 - val_accuracy: 0.9790
Epoch 28/50
63/63 [==============================] - 278s 4s/step - loss: 0.6526 - accuracy: 0.9845 - val_loss: 2.1906 - val_accuracy: 0.9800
Epoch 29/50
63/63 [==============================] - 278s 4s/step - loss: 0.5936 - accuracy: 0.9850 - val_loss: 2.4983 - val_accuracy: 0.9730
Epoch 30/50
63/63 [==============================] - 277s 4s/step - loss: 0.6106 - accuracy: 0.9870 - val_loss: 3.3639 - val_accuracy: 0.9730
Epoch 31/50
63/63 [==============================] - 275s 4s/step - loss: 0.5748 - accuracy: 0.9865 - val_loss: 3.5284 - val_accuracy: 0.9740
Epoch 32/50
63/63 [==============================] - 244s 4s/step - loss: 0.9212 - accuracy: 0.9855 - val_loss: 2.2073 - val_accuracy: 0.9760
Epoch 33/50
63/63 [==============================] - 228s 4s/step - loss: 0.7309 - accuracy: 0.9820 - val_loss: 2.2477 - val_accuracy: 0.9730
Epoch 34/50
63/63 [==============================] - 249s 4s/step - loss: 0.6731 - accuracy: 0.9865 - val_loss: 2.5804 - val_accuracy: 0.9750
Epoch 35/50
63/63 [==============================] - 289s 5s/step - loss: 0.5296 - accuracy: 0.9870 - val_loss: 2.7682 - val_accuracy: 0.9750
Epoch 36/50
63/63 [==============================] - 307s 5s/step - loss: 0.6383 - accuracy: 0.9880 - val_loss: 2.1080 - val_accuracy: 0.9750
Epoch 37/50
63/63 [==============================] - 309s 5s/step - loss: 0.7419 - accuracy: 0.9865 - val_loss: 3.5024 - val_accuracy: 0.9720
Epoch 38/50
63/63 [==============================] - 311s 5s/step - loss: 0.5847 - accuracy: 0.9875 - val_loss: 3.0731 - val_accuracy: 0.9700
Epoch 39/50
63/63 [==============================] - 307s 5s/step - loss: 0.4453 - accuracy: 0.9915 - val_loss: 3.1461 - val_accuracy: 0.9720
Epoch 40/50
63/63 [==============================] - 311s 5s/step - loss: 0.6556 - accuracy: 0.9865 - val_loss: 1.8087 - val_accuracy: 0.9800
Epoch 41/50
63/63 [==============================] - 311s 5s/step - loss: 0.5968 - accuracy: 0.9850 - val_loss: 1.5588 - val_accuracy: 0.9840
Epoch 42/50
63/63 [==============================] - 308s 5s/step - loss: 0.5385 - accuracy: 0.9860 - val_loss: 1.4954 - val_accuracy: 0.9860
Epoch 43/50
63/63 [==============================] - 312s 5s/step - loss: 0.5786 - accuracy: 0.9870 - val_loss: 2.1804 - val_accuracy: 0.9760
Epoch 44/50
63/63 [==============================] - 307s 5s/step - loss: 0.6282 - accuracy: 0.9905 - val_loss: 2.0469 - val_accuracy: 0.9780
Epoch 45/50
63/63 [==============================] - 311s 5s/step - loss: 0.3377 - accuracy: 0.9910 - val_loss: 1.9917 - val_accuracy: 0.9740
Epoch 46/50
63/63 [==============================] - 315s 5s/step - loss: 0.6252 - accuracy: 0.9855 - val_loss: 3.7782 - val_accuracy: 0.9610
Epoch 47/50
63/63 [==============================] - 312s 5s/step - loss: 0.4625 - accuracy: 0.9875 - val_loss: 2.7104 - val_accuracy: 0.9750
Epoch 48/50
63/63 [==============================] - 310s 5s/step - loss: 0.4728 - accuracy: 0.9855 - val_loss: 2.2380 - val_accuracy: 0.9740
Epoch 49/50
63/63 [==============================] - 311s 5s/step - loss: 0.2794 - accuracy: 0.9925 - val_loss: 2.5046 - val_accuracy: 0.9720
Epoch 50/50
63/63 [==============================] - 314s 5s/step - loss: 0.5221 - accuracy: 0.9885 - val_loss: 1.8299 - val_accuracy: 0.9770
In [ ]:
acc = history_2.history["accuracy"]
val_acc = history_2.history["val_accuracy"]
loss = history_2.history["loss"]
val_loss = history_2.history["val_loss"]
epochs = range(1, len(acc) + 1)
plt.plot(epochs, acc, "bo", label="Training accuracy")
plt.plot(epochs, val_acc, "b", label="Validation accuracy")
plt.title("Training and validation accuracy")
plt.legend()
plt.figure()
plt.plot(epochs, loss, "bo", label="Training loss")
plt.plot(epochs, val_loss, "b", label="Validation loss")
plt.title("Training and validation loss")
plt.legend()
plt.show()

Evaluate the model with test set¶

In [ ]:
test_model_2 = keras.models.load_model(
    "best_model.hdf5")
test_loss_2, test_acc_2 = test_model_2.evaluate(test_dataset)
print(f"Test accuracy: {test_acc_2:.3f}")

63/63 [==============================] - 202s 3s/step - loss: 2.4849 - accuracy: 0.9765
Test accuracy: 0.976
In [ ]:
history_2_dict = history_2.history
hist_df_2 = pd.DataFrame(history_2_dict)
hist_df_2
Out[ ]:
loss accuracy val_loss val_accuracy
0 17.168463 0.8895 4.984055 0.965
1 6.100548 0.9440 3.513541 0.974
2 4.568997 0.9555 7.937505 0.956
3 4.880500 0.9600 3.787812 0.978
4 4.338428 0.9570 3.422342 0.974
5 3.038173 0.9710 5.025804 0.970
6 4.658338 0.9675 7.039078 0.963
7 2.984732 0.9710 3.860712 0.972
8 3.273030 0.9730 4.545559 0.970
9 1.804847 0.9785 4.796463 0.972
10 2.569045 0.9740 2.521694 0.976
11 1.796767 0.9795 3.094885 0.975
12 2.565411 0.9725 5.085546 0.963
13 1.623403 0.9785 4.557412 0.966
14 1.942382 0.9750 2.981691 0.979
15 1.570130 0.9820 2.940545 0.977
16 0.935390 0.9845 4.168601 0.976
17 1.073169 0.9855 3.529336 0.977
18 1.669256 0.9810 3.367442 0.974
19 1.327880 0.9820 2.747477 0.975
20 1.328078 0.9800 2.441870 0.975
21 1.010675 0.9845 2.462678 0.976
22 0.684676 0.9865 5.768030 0.955
23 1.097286 0.9825 2.898780 0.976
24 1.143521 0.9825 2.711579 0.975
25 0.862628 0.9800 2.416454 0.977
26 0.859436 0.9845 2.258896 0.979
27 0.652574 0.9845 2.190644 0.980
28 0.593573 0.9850 2.498260 0.973
29 0.610646 0.9870 3.363908 0.973
30 0.574804 0.9865 3.528352 0.974
31 0.921167 0.9855 2.207339 0.976
32 0.730911 0.9820 2.247656 0.973
33 0.673076 0.9865 2.580405 0.975
34 0.529604 0.9870 2.768208 0.975
35 0.638289 0.9880 2.107959 0.975
36 0.741914 0.9865 3.502442 0.972
37 0.584667 0.9875 3.073118 0.970
38 0.445315 0.9915 3.146147 0.972
39 0.655612 0.9865 1.808740 0.980
40 0.596796 0.9850 1.558759 0.984
41 0.538528 0.9860 1.495353 0.986
42 0.578584 0.9870 2.180365 0.976
43 0.628239 0.9905 2.046947 0.978
44 0.337693 0.9910 1.991745 0.974
45 0.625182 0.9855 3.778216 0.961
46 0.462508 0.9875 2.710402 0.975
47 0.472815 0.9855 2.237985 0.974
48 0.279399 0.9925 2.504649 0.972
49 0.522063 0.9885 1.829930 0.977

Conclusions¶

Training & Validation loss

  • This model performs well on the training data and validation data.
  • This model shows a best fit.
  • The gap between validation and training loss is very low.
  • This is an indicator of best fit model.

Accuracy

  • We are getting good accuracy for training set and validation set.
  • The accuracy on test set is .976.
  • This means that 97.6 % of the images are correctly predicted by the model.

Weights

  • The best set of weights seems to correspond to epoch 35, where val_accuracy is the highest (0.981) and val_loss is relatively low compared to other epochs (1.629161).

Summarised Conclusion is give below

Model Name Highest Val Accuracy Val_Loss Epoch
Model_1 0.741 0.720664 13
Model_2 Fine Tuning 0.981 1.629161 35

Accuracy for test set

Model Name Accuracy
Model_1 0.729
Model_2 Fine Tuning 0.976

Relative performance of the models¶

I am using ImageDatagenerator in this step to get lables and features for validation dataset.

In [ ]:
import os
import numpy as np
from keras.preprocessing.image import ImageDataGenerator
from keras.models import load_model
from sklearn.metrics import accuracy_score, confusion_matrix, classification_report, precision_recall_curve
import matplotlib.pyplot as plt

validation_dir = 'datacats_and_dogs_small/validation'

img_height, img_width = 180, 180  
batch_size = 32  
validation_image_generator = ImageDataGenerator(rescale=1./255)

val_data_gen   = validation_image_generator.flow_from_directory(
                    validation_dir,
                    target_size=(img_height, img_width),
                    batch_size=batch_size,
                    class_mode='binary')

Found 1000 images belonging to 2 classes.

Model 1¶

Accuracy

In [ ]:
# Load the best model
model1_new = load_model('best_model_model1.hdf5')
y_pred_model1 = model1_new.predict(val_data_gen)
y_pred_class_model1 = np.round(y_pred_model1).astype(int)
validation_labels = val_data_gen.classes

accuracy_model1 = accuracy_score(validation_labels, y_pred_class_model1)
print("Accuracy - Model 1:", accuracy_model1)

32/32 [==============================] - 6s 184ms/step
Accuracy - Model 1: 0.5

Confusion Matrix

In [ ]:
# Confusion matrix
conf_matrix_model1 = confusion_matrix(validation_labels, y_pred_class_model1)
sns.heatmap(conf_matrix_model1, annot=True, fmt='d', cmap='Blues')
Out[ ]:
<Axes: >

Precision, Recall, F1-Score

In [ ]:
# Generate classification report
report = classification_report(validation_labels, y_pred_class_model1, output_dict=True)

# Convert report to pandas DataFrame
report_df = pd.DataFrame(report).transpose()

# Display the report
display(report_df)
precision recall f1-score support
0 0.50 1.0 0.666667 500.0
1 0.00 0.0 0.000000 500.0
accuracy 0.50 0.5 0.500000 0.5
macro avg 0.25 0.5 0.333333 1000.0
weighted avg 0.25 0.5 0.333333 1000.0

Precision-Recall Curve

In [ ]:
# Precision-Recall curve
precision_model1, recall_model1, _ = precision_recall_curve(validation_labels, y_pred_model1)
plt.plot(recall_model1, precision_model1, label='Model 1')
plt.xlabel('Recall')
plt.ylabel('Precision')
plt.title('Precision-Recall Curve')
plt.legend()
plt.show()

Model 2¶

Accuracy

In [ ]:
# Load the best model
model2_new = load_model('best_model.hdf5')
y_pred_model2 = model2_new.predict(val_data_gen)
y_pred_class_model2 = np.round(y_pred_model2).astype(int)

accuracy_model2 = accuracy_score(validation_labels, y_pred_class_model2)
print("Accuracy - Model 2:", accuracy_model2)

32/32 [==============================] - 101s 3s/step
Accuracy - Model 2: 0.5

Confusion matrix

In [ ]:
# Confusion matrix
conf_matrix_model2 = confusion_matrix(validation_labels, y_pred_class_model2)
sns.heatmap(conf_matrix_model2, annot=True, fmt='d', cmap='Blues')
Out[ ]:
<Axes: >
In [ ]:
# Generate classification report
report_2 = classification_report(validation_labels, y_pred_class_model2, output_dict=True)

# Convert report to pandas DataFrame
report_df_2 = pd.DataFrame(report_2).transpose()

# Display the report
display(report_df_2)
precision recall f1-score support
0 0.50 1.0 0.666667 500.0
1 0.00 0.0 0.000000 500.0
accuracy 0.50 0.5 0.500000 0.5
macro avg 0.25 0.5 0.333333 1000.0
weighted avg 0.25 0.5 0.333333 1000.0
In [ ]:
# Precision-Recall curve
precision_model2, recall_model2, _ = precision_recall_curve(validation_labels, y_pred_model2)
plt.plot(recall_model2, precision_model2, label='Model 2')
plt.xlabel('Recall')
plt.ylabel('Precision')
plt.title('Precision-Recall Curve')
plt.legend()
plt.show()

Conclusions¶

  • Fine tuning with vgg16 has good accuracy in test set.
  • It has a accuracy of .976 fo test set.
  • That means the model is able to predict 97.6 % of images correctly.
  • The precision recall curve for model 2 is closer to top left corner of the graph than model 1 which means for a given level of recall model 2 has high precision.
  • So Model 2 (vgg16) is the good model.